Adaptive structured parallelism for distributed heterogeneous architectures: a methodological approach with pipelines and farms

González–Vélez, Horacio; Cole, Murray

doi:10.1002/cpe.1549

Cited by 14 publications

(5 citation statements)

References 40 publications

(34 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Murray Cole's group at the School of Informatics of the University of Edinburgh studied the implications of using skeletons in the scheduling algorithms by using models based on process algebras (specifically PEPA in [7,15]). González-Vélez and Cole [17] presented a methodology that uses the information obtained from an application based on a skeleton to improve performance at run-time. It was based on heterogeneous distributed systems and considered task farm and pipeline patterns.…”

Section: Related Workmentioning

confidence: 99%

HeDPM: load balancing of linear pipeline applications on heterogeneous systems

et al. 2017

View full text Add to dashboard Cite

This work presents a new algorithm, called Heterogeneous Dynamic Pipeline Mapping, that allows for dynamically improving the performance of pipeline applications running on heterogeneous systems. It is aimed at balancing the application load by determining the best replication (of slow stages) and gathering (of fast stages) combination taking into account processors computation and communication capacities. In addition, the algorithm has been designed with the requirement of keeping complexity low to allow its usage in a dynamic tuning tool. For this reason, it uses an analytical performance model of pipeline applications that addresses hardware heterogeneity and which depends on parameters that can be known in advance or This work has been partially supported by Ministerio de Economía y Competitividad MINECO-Spain under contract TIN2014-53234-C2-1-R and Generalitat de Catalunya GenCat-DIUiE (GRR) 2014-SGR-576.

show abstract

Section: Related Workmentioning

confidence: 99%

HeDPM: load balancing of linear pipeline applications on heterogeneous systems

et al. 2017

View full text Add to dashboard Cite

show abstract

“…A methodology for partitioning divisible loads for the farm pattern in heterogeneous CPU-only clusters has been proposed in APSARA [7]. It uses static and dynamic calibration techniques to calculate the execution speeds of the worker processors.…”

Section: Related Workmentioning

confidence: 99%

On the Dynamic Scheduling of Task Farm Patterns on a Heterogeneous CPU-GPGPU Environment

Zhang

Goswami

Goodarzi

2008

Proceedings of the 2014 International C* Conference on Computer Science &Amp; Software Engineering - C3S2E '14

View full text Add to dashboard Cite

Heterogeneous clusters and multi-core environments are gradually surpassing the homogeneous systems due to their high performance and flexibility. Task scheduling in these systems is an extensively studied subject. However, in a heterogeneous architecture consisting of multi-core CPUs and many-core GPGPUs (General Purpose Graphics Processor Units), task mapping becomes much more complex due to differences in architectures and programming models among the processors. Consequently, designing a scheduler which facilities a balanced distribution of loads by taking full advantage of the processing power of a CPU-GPGPU system is nontrivial. In this paper we discuss a multi-round scheduling algorithm and a scheduling framework for farm-pattern based applications on such a system. This is an important step towards designing a full-scale patternbased scheduler to automatically and efficiently map the parallel tasks to the heterogeneous processors. As a proof of concept, we have designed a scheduling framework for the task-farm pattern based applications. The framework provides the necessary "separation of concerns" and hides the underlying complex scheduling details from the programmer. The experimental results demonstrate that our dynamic scheduling algorithm achieves better to similar performances as compared to some of the wellknown scheduling algorithms for CPU-GPGPU systems.

show abstract

“…The second solution that we would like to highlight is part of the ASPARA project [13] leaded by Murray Cole and Horacio Gónzales-Vélez. This work proposes a methodology to introduce adaptability in skeletons.…”

Section: Related Workmentioning

confidence: 99%

Self-Configuration and Self-Optimization Autonomic Skeletons using Events

Pabón¹,

Henrio

2014

Proceedings of Programming Models and Applications on Multicores and Manycores

View full text Add to dashboard Cite

This paper presents a novel way to introduce self-configuration and self-optimization autonomic characteristics to algorithmic skeletons using event driven programming techniques. Based on an algorithmic skeleton language, we show that the use of events greatly improves the estimation of the remaining computation time for skeleton execution. Events allow us to precisely monitor the status of the execution of algorithmic skeletons. Using such events, we provide a framework for the execution of skeletons with a very high level of adaptability. We focus mainly on guaranteeing a given execution time for a skeleton, by optimizing autonomically the number of threads allocated. The proposed solution is independent from the platform chosen for executing the skeleton for example we illustrate our approach in a multicore setting, but it could also be adapted to a distributed execution environment.

show abstract

Adaptive structured parallelism for distributed heterogeneous architectures: a methodological approach with pipelines and farms

Cited by 14 publications

References 40 publications

HeDPM: load balancing of linear pipeline applications on heterogeneous systems

HeDPM: load balancing of linear pipeline applications on heterogeneous systems

On the Dynamic Scheduling of Task Farm Patterns on a Heterogeneous CPU-GPGPU Environment

Self-Configuration and Self-Optimization Autonomic Skeletons using Events

Contact Info

Product

Resources

About